Enzyme-catalyzed resolution of 3,8-dioxatricyclo[3.2.1.02,4]octane-6-carboxylic esters and the application to the synthesis of 3-epishikimic acid

Article abstract of DOI:10.1016/j.tetasy.2009.07.049

3,8-Dioxatricyclo[3.2.1.0^2,4]octane-6-carboxylic acid, whose racemic form is readily available on a large scale, is a versatile starting material for the synthesis of carbasugars and carbocyclic biologically active natural products. In this stu

Full text of DOI:10.1016/j.tetasy.2009.07.049

Tetrahedron: Asymmetry 20 (2009) 2105–2111
Contents lists available at ScienceDirect
Tetrahedron: Asymmetry
journal homepage: www.elsevier.com/locate/tetasy
Enzyme-catalyzed resolution of 3,8-dioxatricyclo[3.2.1.02,4]octane-6-carboxylic
esters and the application to the synthesis of 3-epishikimic acid
Manabu Hamada a, Yoshikazu Inami b, Yasuhito Nagai c, Toshinori Higashi b, Mitsuru Shoji b,
Seiichiro Ogawa d, Kazuo Umezawa a, Takeshi Sugai b,
*
a Department of Applied Chemistry, Keio University, 3-14-1 Hiyoshi, Yokohama 223-8522, Japan
b Faculty of Pharmacy, Keio University, 1-5-30, Shibakoen, Minato-ku, Tokyo 105-8512, Japan
c Department of Chemistry, Keio University, 3-14-1 Hiyoshi, Yokohama 223-8522, Japan
d Department of Bioscience and Bioinformatics, Keio University, 3-14-1 Hiyoshi, Yokohama 223-8522, Japan
a r t i c l e i n f o
a b s t r a c t
3,8-Dioxatricyclo[3.2.1.02,4]octane-6-carboxylic acid, whose racemic form is readily available on a large
scale, is a versatile starting material for the synthesis of carbasugars and carbocyclic biologically active
natural products. In this study, the enzyme-catalyzed kinetic resolution was attempted on a variety of
corresponding carboxylic esters. The hydrophobic and hydrophilic properties of ester substituents greatly
affected the rate of reaction and the enantioselectivity. Hydrolysis of the corresponding 20-chloroethyl
ester with pig liver esterase worked well in a highly enantioselective manner (E = 116) to give the hydro-
lyzate (90.6% ee) and unreacted ester recovery (99.4% ee). The hydrolyzate is a precursor for (ꢀ)-oseltam-
ivir phosphate, and a route to (3S,4S,5R)-(ꢀ)-3-epishikimic acid was developed from the recovered ester.
Ó 2009 Elsevier Ltd. All rights reserved.
Article history:
Received 14 June 2009
Accepted 6 July 2009
Available online 6 October 2009
1. Introduction
2. Results and discussion
3,8-endo-Dioxatricyclo[3.2.1.02,4]octane-6-carboxylate (acid or
ester) 1 is a polyoxygenated cyclohexenecarboxylate with many
controlled stereocenters which has been developed as a starting
material for carbasugars, including carba-Neu5Ac and carba-KDO
as shown in Scheme 1.1 In the synthesis of epoxyquinols and re-
lated compounds, Hayashi et al. have explored new entries. An effi-
cient selective ring-opening reaction was a key step to provide
cyclohexenecarboxylate.2,3 The racemic form of 1 is available in
large quantity.1,6 So far, however, the supply of the enantiomeri-
cally pure form has been rather limited, except for asymmetric
Diels–Alder reactions controlled by chiral auxiliary4,5 or tedious
preferential crystallization of the diastereomeric salts of a certain
precursor.6
Herein, we report the kinetic resolution by the action of
hydrolytic enzymes on esters 1b–g (Scheme 1). The hydrolysis of
carboxylic esters 1 has a clear and significant advantage, for the
large-scale enantiomeric resolution. In this approach, the hydroly-
zates, carboxylic acids, and the unreacted recoveries, the esters, are
separable by only extractive workup under properly adjusted pH
conditions. If these attempts are successful, the unreacted recover-
ies, the esters would be converted to (3S,4S,5R)-(ꢀ)-3-epishikimic
acid 2.
The preparation of racemic esters 1b–g is shown in Scheme 2.
The Diels–Alder reaction between furan and acrylic acid, and the
subsequent iodolactonization of the crude endo–exo mixture
(8:2) yielded racemic iodolactone 5.1 The hydrolysis of the lactone
followed by the direct alkylation of the resulting carboxylate pro-
vided esters 1b–g.
Our first candidate was the ethyl ester 1b, as it showed a lower
melting point (mp 52–53 °C) than the corresponding methyl ester
1c (mp 75 °C). The lower melting point of the crystalline substrate
has been observed to be advantageous for enzyme-catalyzed
hydrolysis under aqueous conditions.7 So far, attempts at the ki-
netic resolution of the similar endo-carboxylic acid 3, by the action
of hydrolytic enzymes on the corresponding ester, have only re-
sulted in moderate enantioselectivity.8
Our substrate 1b reacted poorly with many kinds of hydrolytic
enzymes, probably due to high steric hindrance around the
endo-oriented ester moiety. Among the proteases (Rhizopus sp.,
XP-415; A. melleus, XP-488, Nagase ChemteX Co.), lipases (Candida
rugosa, Meito OF; Candida antarctica, Roche diagnostics, Chirazyme
L-2), and esterases (Klebsiella oxytoca, SNSM-87, Nagase ChemteX
Co.; pig liver esterase, Sigma), only pig liver esterase (PLE) showed
substantial hydrolysis (Scheme 3). However, the conversion and
enantioselectivity (E-value) were as low as 33% and 11%, respec-
tively, as shown in entry 1, Table 1.
The decrease in bulkiness from ethyl to methyl 1c (entry 2)
only resulted in lower selectivity (E = 7). We then introduced an
* Corresponding author. Tel./fax: +81 3 5400 2665.
E-mail address: sugai-tk@pha.keio.ac.jp (T. Sugai).
0957-4166/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2009.07.049

2106
M. Hamada et al. / Tetrahedron: Asymmetry 20 (2009) 2105–2111
O
O
O
O
5
4
6
2
+
+
1
CO2R
CO2R
O
O
CO2R
O
O
CO2R
( )-1b-g
(-)-1b-g
(+)-1b-g
hydrolysis
O
O
O
O
enzyme-catalyzed
kinetic resolution
a)
+
+
CO2R
(-)-1b-g
CO2R
O
CO2H
O
O
CO2H
O
(+)-1a
(-)-1b-g
(+)-1a
Scheme 3. Enzyme-catalyzed hydrolysis of endo-epoxyesters. Reagents and condi-
tions: (a) pig liver esterase (Sigma E2884), rt, 24 h.
HO
HO
OH
OH
HO
HO
AcHN
HO
CO2H
OH
HO
OH
CO2H
Table 1
PLE-catalyzed hydrolysis of ( )-1b–ga
carba-Neu-5-Ac
O
carba-KDO
Entry
Substrate
R
Conv. (%)
E-value
O
1
2
3
4
5
6
1b
1c
1d
1e
1f
CH2CH3
CH3
CH2Cl
CH2CH2Cl
CH2CONH2
CH2CF3
33
32
NA
51
26
50
11
7
NA
78
5
OH
HO
O
O
O
OH
O
O
OH
H
1g
9
O
a
RKTS-34
For reaction conditions and evaluation of enantioselectivity, see Section 4.
epoxyquinol A
3
HO
CO2H
to directly replace one hydrogen atom with chlorine 1d (entry 3)
turned out to be unsuccessful, because 1d was too unstable. The
introduction of 2-chloroethyl9 (1e, entry 4) and carbamylmethyl10
1f (entry 5) groups brought about contrasting effects. The former
showed a great increase in selectivity (E = 78) and the conversion
reached nearly 50%.
This result suggested that the reactivity of the ‘fast’ isomer was
enhanced, this is well supported by applying an empirical model of
the catalytic site of PLE as proposed by Jones (Fig. 1, top).11 The tri-
cyclic skeleton of the substrate occupies a large hydrophobic pock-
et (L). In turn, the 2-chloroethyl group fits a small hydrophobic site
(S) of the model so that the attack by the hydroxyl group of a serine
residue easily takes place. The 1H NMR spectra of 1e prompted us
to hypothesize that an extended conformation is advantageous to
fit in the hydrophobic site of PLE. The two geminal protons of Ha
and Hb (Fig. 1, top) were not equivalent (Ha: 4.44 ppm, Hb:
HO
OH
(3S,4S,5R)-2
3-epishikimic acid
Scheme 1. Enzyme-catalyzed kinetic resolution of 3,8-dioxatricyclo[3.2.1.02,4]-
octane-6-carboxylic ester 1 and utilization thereof.
O
O
a)
+
+
O
CO2H
CO2H
CO2H
endo-3
exo-4
4.32 ppm,
Dd 0.12 ppm), and one of them would be located in
8 : 2
the eclipsed position of the ester carbonyl group make the chloro-
ethyl group fit into the hydrophobic site (S) (Fig. 1, top).
On the other hand, the lower selectivity in the carbamylmethyl
ester 1f (entry 5, E = 5), as well as the lower conversion, means that
the ‘fast’ isomer is not preferable in an enzyme-catalyzed reaction.
The 1H NMR measurement suggested that 1f has a different confor-
O
O
7
3
6
b)
c)
I
( )-3
1
2
CO2R
O
O
O
mation from 1e, as judged by a smaller
Hd (Fig. 1, bottom) (Hc: 4.69 ppm, Hd: 4.62 ppm,
D
d value between Hc and
d 0.07 ppm).
( )-5
( )-1b-g
D
When 1f in this conformation occupies the catalytic site, a car-
bamylmethyl group fits the hydrophilic site, while the ester car-
bonyl group is far away from the nucleophilic serine hydroxyl
group (Fig. 1, bottom). The crooked conformation of 1f is supported
by the fact that 1f is prone to epimerization to an exo-ester, when
under treatment with base followed by an intramolecular delivery
of protons from CONH2 group. In contrast, b-elimination exclu-
sively occurs very quickly in 1e even under low temperatures, as
described later.
Scheme 2. Preparation of endo-epoxy esters. Reagents and conditions: (a) see Ref.
6. The ratio between endo- and exo-adducts (8:2) was determined by 1H NMR; (b)
NaHCO3, H2O/I2, THF, rt, 24 h (quant.); (c) KOH, DMF, 40 °C, 24 h/RI, 50 °C, 6 h
(yields: see Section 4).
electron-withdrawing group on the ester moiety, expecting that
the rate of the ‘fast’ (1S,2R,4S,5R,6S)-isomer would be enhanced,
so that the enantioselectivity becomes higher. Our first attempt

M. Hamada et al. / Tetrahedron: Asymmetry 20 (2009) 2105–2111
2107
hydrolyzate, followed by b-elimination2 of the resulting ester 1c
using LHMDS gave (ꢀ)-6a. The absolute configuration of this sam-
ple was unambiguously determined to be (1S,5S,6R) by the minus
sign of the rotation value compared with literature data.2 This re-
sult is consistent with the stereochemical preference of the ‘fast’
isomer being similar to that in the case of the ester of endo-3
(Scheme 2).8
Since the combination of highly enantioselective catalyst and
substrate was optimized, the next task was the establishment of
suitable reaction conditions for the large-scale preparation of ( )-
1e from iodolactone ( )-5. The conventional method for ester syn-
thesis is the hydrolysis of iodolactone to give the salt of acid 1a
with concomitant epoxide ring formation, and subsequent treat-
ment with an alkylating agent in one pot (Scheme 5).
hydrophobic
site
(L)
O
H
hydrophobic
site (S)
Ha
O
C
O
O
O
C
CH2Cl
Hb
Ser
eclipsed
1e: fast isomer
O
O
O
O
1f: fast isomer
I
c)
a,b)
O
C
O
O
Hc
O
CO2CH2CH2Cl
O
CO2H
( )-1a
O
C
H2N C
O
O
O
Hd
( )-5
1e
( )-
hydrophilic
site
O
H2N
hydrophobic
site (L)
Hc
O
I
I
Hd
O
O
O
O
OCH2CH2Cl
O
C OCH2CH2Cl
O
O
H
O
long distance
Ser
O
( )-5
( )-1e
H
Scheme 5. Preparation of 2-chloroethyl ester 1e. Reagents and conditions: (a) KOH,
EtOH, 70 °C, 5 h; (b) 1 M HCl (quant.); (c) ClCH2CH2X, base, additive. For the
conditions, see Table 2.
Figure 1. Behavior of 2-chloroethyl 1e and carbamylmethyl 1f esters in Jones’
empirical catalytic site model of PLE.
Based on these observations, the more electron-withdrawing
and hydrophobic trifluoromethylester 1g was prepared with the
aim of obtaining improved results. However, 1g was again substan-
tially unstable even in neutral buffer solution. The lower enanti-
oselectivity (E = 9) was probably due to enzyme-uncatalyzed
spontaneous hydrolysis.
Esterification with 1-chloro-2-iodoethane smoothly proceeded
to give 1e in 80% yield (Table 2, entry 1). However, this mixed
dihalide is quite expensive and not suitable for large-scale prepara-
tion. The replacement of this iodide with inexpensive 1,2-dichloro-
ethane resulted in only a 33% yield (Table 2, entry 2). Next, KI
(4 equiv) was added with the expectation of the in situ formation
of iodide, which caused the yield to drop to 3% (Table 2, entry 3).
By analysis of the by-products, iodolactone 5 appeared again dur-
ing the alkylation, even after confirmation of completion of the
hydrolysis in the first step (Scheme 5 and 5?1a). At this stage
we became aware that the presence of iodide ions in the reaction
mixture has a deleterious effect, as the direct attack of an iodide
ion opens the epoxide ring of 2-chloroethyl ester itself and the
reaction reaches an equilibrium between iodolatone 5 and 2-chlo-
roethyl ester 1e (Scheme 5, bottom). Thus we switched the proce-
dure into the following conditions. Acid 1a was extracted after
acidic workup, and then it was separately incubated with 1,2-
The absolute configuration of the ‘fast’ isomer in 1e was deter-
mined as follows in Scheme 4. Esterification of (+)-acid 1a, the
O
O
a)
O
CO2H
O
CO2CH3
(+)-1a
(+)-1c
OH
Table 2
b)
Alkylation of free carboxylic acid ( )-1a
O
Entry
Base
Alkylating agent
X (equiv)
Additive
Yield (%)
CO2CH3
1
2
3
4
KOH
KOH
KOH
K2CO3
I (1.5)
None
None
KI
80
33
3
(-)-6a
Cl (3.0)
Cl (3.0)
Cl (4.0)
Scheme 4. Absolute configuration of (+)-1a. Reagents and conditions: (a) Cs2CO3,
MeI, DMF, 50 °C, 24 h (59%); (b) LHMDS, THF, ꢀ78 °C, 1 h (70%).
None
71

2108
M. Hamada et al. / Tetrahedron: Asymmetry 20 (2009) 2105–2111
dichloroethane and K2CO3. Under these iodide-free conditions, the
reaction was successful and a 71% yield was recorded (Table 2, en-
try 4). In the case of scaled-up conditions, the iodide-free potas-
sium salt of 1a became accessible via an alternative method by
careful observation during hydrolysis. When iodolactone 5 was
hydrolyzed in ethanol, first, a precipitate appeared. This was pro-
ven by NMR analysis to be a potassium salt of iodohydrin carbox-
ylic, an intermediate acid. If the hydrolysis was continued under
prolonged heating at 70 °C, the precipitate disappeared. The result-
ing potassium salt and all of the by-products were soluble in eth-
anol. Then the reaction mixture was dried onto silica gel, and
elution with methanol provided the desired carboxylate salt. Most
of the contaminant, especially inorganic salts such as KI, was re-
moved by adsorption on silica gel.
O
O
+
CO2CH2CH2Cl
O
CO2CH2CH2Cl
(+)-1e
O
(-)-1e
O
O
a)
+
CO2CH2CH2Cl
(-)-1e
O
CO2H
O
(+)-1a
The PLE catalyzed hydrolysis in the scaled-up experiment
proceeded in a reproducible manner, and (+)-1a (54.7% yield) and
(ꢀ)-1e (42.7% yield) were obtained with only an extractive
work-up separation. In this case, both ees of (+)-1a (90.6% ee)
and (ꢀ)-1e (99.4% ee) were unambiguously determined (see
Section 4) (E = 116). The simple recrystallization of acid (+)-1a
from EtOAc enhanced the enantiomeric excess to 96.4%. It is note-
worthy that acid (+)-1a has been reported as the starting material
for oseltamivir phosphate by Terashima and Ujita.12
CO2CH2CH2Cl
b)
O
(-)-1e
OH
(+)-6b
Our product, epoxyester 1e, has the same level of oxygen func-
tionality in suitable positions as shikimic acid and related com-
pounds. Naturally occurring shikimic acid (3R,4S,5R)-7 is
essential for the industrial synthesis of (ꢀ)-oseltamivir phos-
phate.13,14 The process for fermentative production has already
been established,15,16 as (3R,4S,5R)-7 is a biosynthetic key interme-
diate for the well-known shikimate pathway. An epimeric form,
(3S,4S,5R)-2 (3-epishikimic acid), has recently gained attention as
the starting material for vitamin D3 precursors,17 the synthon of
natural products,18,19 and as a template for combinatorial synthe-
sis.20 The availability of this epi-form, however, is very low due
to no direct biosynthetic pathway. An attempted acid-catalyzed
epimerization under harsh conditions only results in a stereoiso-
meric mixture with parent (3R,4S,5R)-shikimic acid 7, accompa-
nied with the dehydrated 4-hydroxybenzoic acid.18 Otherwise, a
tedious multi-step conversion was required, involving selective
protection of 4,5-trans diol and inversion at C-3.19
HO
CO2H
HO
HO
CO2H
c,d)
HO
OH
OH
(3R,4S,5R)-7
shikimic acid
(3S,4S,5R)-2
Scheme 6. Derivation of hydrolyzate to (3S,4S,5R)-3-epishikimic acid 2a. Reagents
and conditions: (a) pig liver esterase, 0.2 M phosphate buffer (pH 7.0), [42.7% for
(1R,2S,4R,5S,6R)-(ꢀ)-1e (99.4% ee)], [54.7% for (1S,2R,4S,5R,6S)-(+)-1a (90.6% ee)];
(b) LHMDS, THF, ꢀ78 °C, 1 h, (96%); (c) KOH, THF, H2O, 50 °C, 1 h; (d) TFA, H2O,
50 °C, 3 h (79.7%).
4. Experimental
The above-mentioned situation prompted us to establish a
route from (ꢀ)-1e to epishikimate (3S,4S,5R)-2. Epoxy ester
(ꢀ)-1e was submitted to LHMDS-mediated b-elimination2 to give
(+)-6b (Scheme 6). It is noteworthy that the desired reaction
occurred in as high as 94% yield without any damage on the 2-chlo-
roethyl ester, which would also suffer from b-elimination. For the
transformation of 6b to (3S,4S,5R)-2, the electron-withdrawing
property of the chloroethyl group was quite advantageous. The alka-
line hydrolysis of 6b proceeded very smoothly, and the following
stereoselective epoxide ring opening21 gave (3S,4S,5R)-2a (79.7%);
4.1. Materials and methods
Merck Silica Gel 60 F254 thin-layer plates (1.05744, 0.5 mm
thickness) and Silica Gel 60 (spherical and neutral; 100–210 lm,
37560-79) from Kanto Chemical Co. were used for preparative
thin-layer chromatography and column chromatography, respec-
tively. The commercial PLE preparation was purchased from Sigma.
4.2. Analytical methods
½
a 2D5
ꢁ
¼ ꢀ33:1 (c 0.34, H2O) {lit.21
[
a]
D = ꢀ31.0 (c 0.1, H2O)}; whose
spectroscopic data coincided with those reported previously.21
All melting points are uncorrected. IR spectra were measured as
thin films for oils or ATR for solid on a Jeol FT-IR SPX60 spectrom-
eter. 1H NMR spectra were measured in CDCl3 or D2O at 270 MHz
on a Jeol JNM EX-270 or at 400 MHz on a Jeol JNM GX-400 spec-
trometer or at 400 MHz on a VARIAN 400-MR spectrometer, and
13C NMR spectra were measured in CDCl3 or D2O at 100 MHz on
a Jeol GX-400 spectrometer or at 100 MHz on a VARIAN 400-MR
spectrometer. High resolution mass spectra were recorded on a
Jeol JMS-700 MStation spectrometer. HPLC data were recorded
on Jasco MD-2010 multi-channel detectors and SHIMADZU SPD-
M20A diode array detector. Optical rotation values were recorded
on a Jasco P-1010 polarimeter. Silica Gel 60 (spherical, 100–
3. Conclusion
Based on the pig liver esterase-catalyzed kinetic resolution of
2-chloroethyl 3,8-dioxatricyclo[3.2.1.02,4]octane-6-carboxylate 1e,
an expeditious route for polyhydroxylated cyclohexenoids has been
established. The design of the substrate structure was supported by
conformational analysis and fitness in an enzyme catalytic site
model. The reaction conditions for the synthesis of optimized
substrate, excluding the formation of by-products necessary to
simplify the workup procedure, which is indispensable for prepara-
tive-scale synthesis have been elucidated. 3-Epishikimic acid
should be a more promising starting material in organic synthesis
following our establishment of a scalable supply.
210 lm, 37558-79) of Kanto Chemical Co. was used for column
chromatography. Preparative TLC was performed with E. Merck Sil-
ica Gel 60 F256 plates (0.5 mm thickness, No. 5744).

M. Hamada et al. / Tetrahedron: Asymmetry 20 (2009) 2105–2111
2109
4.3. Screening of hydrolytic enzymes
concentrated in vacuo. The residue was purified by silica gel col-
umn chromatography with hexane–EtOAc (2:1) to afford ethyl es-
ter ( )-1b (316 mg, 91.8%) as a colorless solid, mp 52–53 °C. Its
NMR spectrum was identical with that reported previously.12
The screening of hydrolytic enzymes were performed as fol-
lows. A 2 mL sample tube was charged with an appropriate amount
of racemic ethyl ester 1b (10.0 mg) and potassium phosphate buf-
fer (0.2 M, 0.25 mL, pH 7.0) at room temperature for 24 h in the
presence of several lipases and protease at an amount of 80–
100 mg/mL of phosphate buffer, in the case of pig liver esterase
0.2 mg/mL of phosphate buffer. The progress of the reaction was
monitored by TLC analysis [silica gel, developed with hexane–
EtOAc (1:1)]. The reaction mixture was quenched with citric acid
to pH 2, and extracted with EtOAc. The combined organic phases
were dried over Na2SO4 and concentrated in vacuo. Among the se-
ven enzymes tested, only pig liver esterase showed the progress of
hydrolysis.
4.3.5. Methyl ( )-3,8-dioxatricyclo[3.2.1.02.4]octane-6-carboxyl-
ate 1c
In a similar manner as described for 1b, a solution of iodolactone
( )-5 (0.80 g, 3.01 mmol) in DMF (10 mL) was treated with KOH
(0.40 g, 7.13 mmol) and MeI (1.28 g, 6.56 mmol) to give methyl ester
( )-1c (390 mg, 78.3%) as a colorless solid; mp 75 °C. Its NMR
spectrum was identical with that reported previously.2
4.3.6. 2-Chloroethyl ( )-3,8-dioxatricyclo[3.2.1.02.4]octane-6-
carboxylate 1e
In a similar manner as described for 1b, a solution of iodolac-
tone ( )-5 (266 mg, 1.00 mmol) in DMF (3 mL) was treated with
KOH (132 mg, 2.35 mmol) and ClCH2CH2I (300 mg, 1.58 mmol),
to give 2-chloroethyl ester ( )-1e (171 mg, 78.5%) as a colorless
oil: 1H NMR (400 MHz, CDCl3): d = 1.97 (1H, ddd, J = 4.9, 11.3,
11.3 Hz, H-7exo), 2.07 (1H, dd, J = 4.9, 11.3 Hz, H-7endo), 2.93 (1H,
dt, J = 4.9, 11.3 Hz, H-6), 3.69 (2H, t, J = 5.7 Hz, CH2Cl), 4.01 (1H,
dd, J = 2.4, 4.3 Hz, H-4), 4.11 (1H, dd, J = 2.2, 4.3 Hz, H-2), 4.32
(1H, dt, J = 5.7, 11.3 Hz, CHHCH2Cl, Ha in Fig. 1), 4.44 (1H, dt,
J = 5.7, 11.3 Hz, CHHCH2Cl, Hb in Fig. 1), 4.50 (1H, dt, J = 2.2,
4.9 Hz, H-5), 4.70 (1H, dt, J = 2.2, 4.9 Hz, H-1); 13C NMR (100
MHz, CDCl3): d = 29.5, 41.8, 44.7, 64.1, 66.6, 66.7, 77.5, 78.2,
171.0; IR mmax 3006, 2962, 2358, 1732, 1449, 1334, 1207, 1176,
883 cmꢀ1. Anal. Calcd for C9H11ClO4: C, 49.44; H, 5.07. Found: C,
49.19; H, 5.04.
4.3.1. ( )-7-endo-Oxabicyclo[2.2.1]hept-5-carboxylic acid 3
The known procedure6 was slightly modified in regard to the
reaction temperature. Furan (250 mL) and acrylic acid (250 mL)
were mixed and kept for 6 days at room temperature, for 17 days
at 4 °C, and then for 28 days at 7 °C. The precipitated solids were
recovered by filtration to give carboxylic acid ( )-3 (41.7 g, en-
do:exo = 8:2) as a colorless solid, mp 95–96 °C, lit.6 mp 97–
100 °C. To the above mentioned filtrate was added furan and ac-
rylic acid and kept for one month at 7 °C to give another crop of
crystal. The NMR spectrum was identical with that reported
previously.1,6
4.3.2. ( )-(1R*,2R*,3R*,6R*,7S*)-2-Iodo-4,8-dioxatricyclo[4.2.1.03.7]-
nonan-5-one 5
To a solution of the acid ( )-3 (20.0 g, 142 mmol) in NaHCO3 aq
solution (300 mL) was added dropwise a solution of I2 (40.0 g,
157 mmol) in THF (80 mL) under ice-cooling, and the mixture
was stirred for 68 h at room temperature. To the mixture was
added saturated Na2S2O3 aq solution, and the precipitates were
collected in filtration to give crude iodolactone ( )-5 (25.1 g). The
filtration was extracted with EtOAc (three times). The organic layer
was washed with brine, dried over Na2SO4 and concentrated in va-
cuo, to give another crop of iodolactone ( )-5 (2.00 g) as the sample
for NMR measurement. Its NMR spectrum was identical with that
reported previously.1 The combined yield of above-mentioned ( )-
5 (27.0 g) was 71%, and this was employed for the next step with-
out further purification.
This ester was also able to be prepared by the action of
ClCH2CH2Cl. To a solution of the acid ( )-1a (15.6 mg, 0.10 mmol)
in anhydrous DMF was added K2CO3 (55.2 mg, 0.40 mmol) and
ClCH2CH2Cl (59.4 mg, 0.60 mmol), and the mixture was stirred at
40 °C for 24 h. The same workup as above provided ( )-1e
(16.5 mg, 75.7%).
4.3.7. Scaled-up and preparative synthesis of 1e
A solution of iodolactone ( )-5 (3.20 g, 12.0 mmol) in EtOH
(20 mL) was added KOH (2.00 g, 35.6 mmol) and the mixture was
stirred for 5 h at 70 °C. After removal of volatile materials in vacuo,
the residue was re-dissolved in MeOH. To the mixture was added
silica gel (50 g), and stirred for 30 min. After concentration in va-
cuo, the residual solid was charged on a glass column, and that
was eluted with ethanol to give carboxylic acid ( )-1a (86.4 mg)
as a colorless solid. Further elution with MeOH afforded potassium
salt ( )-1a (2.40 g).
To a solution of the above potassium salt ( )-1a (2.07 g) in DMF
(10 mL) was added ClCH2CH2Cl (5.28 g, 53.4 mmol), and the mix-
ture was stirred at 60 °C for 24 h. The same workup provided ( )-
1e (1.55 g, 71.1%).
4.3.3. ( )-3,8-Dioxatricyclo[3.2.1.02.4]octane-6-carboxylic acid
1a
To a solution of iodolactone ( )-5 (1.01 g, 3.80 mmol) in DMF
(15 mL) was added KOH (0.54 g, 9.98 mmol) and stirred for 24 h
at room temperature. After removal of water in vacuo, the residue
was added 1 M HCl to pH 2. The mixture was extracted with EtOAc
(10 times), and the combined organic layer was washed with brine,
dried over Na2SO4, and concentrated in vacuo. The residue was
purified by silica gel column chromatography with CHCl3–MeOH
(6:1) to afford carboxylic acid ( )-1a (193 mg, 33.5%) as a colorless
solid, mp 153–154 °C. Its NMR spectrum was identical with that
reported previously.12
4.3.8. Carbamylmethyl ( )-3,8-dioxatricyclo[3.2.1.02.4]octane-6-
carboxylate 1f
In a similar manner as described for 1b, a solution of iodolac-
tone ( )-5 (0.80 g, 3.01 mmol) in DMF (5 mL) was treated with
KOH (0.40 g, 7.13 mmol) and ClCH2CONH2 (0.84 g, 8.98 mmol),
gave carbamylmethyl ester ( )-1f (423 mg, 65.9%) as a colorless so-
lid. Further purification by recrystallization from EtOAc afforded
( )-1f: mp 131.0–133.0 °C: 1H NMR (400 MHz, CDCl3): d = 2.03
(1H, ddd, J = 4.6, 11.2, 11.6 Hz, H-7exo), 2.09 (1H, dd, J = 4.3,
11.6 Hz, H-7endo), 3.02 (1H, dt, J = 4.6, 11.2 Hz, H-6), 4.11 (1H, dd,
J = 1.9, 4.4 Hz, H-4), 4.18 (1H, dd, J = 2.2, 4.4 Hz, H-2), 4.55 (1H,
dt, J = 2.2, 4.6 Hz, H-5), 4.62 (1H, d, J = 15.6 Hz, CHHCONH2, Hc in
Fig. 1), 4.69 (1H, d, J = 15.6 Hz, CHHCONH2, Hd in Fig. 1), 4.77
(1H, dt, J = 1.9, 4.4 Hz, H-1), 5.81 (1H, br s, NH2), 6.69 (1H, br s,
4.3.4. Ethyl ( )-3,8-dioxatricyclo[3.2.1.02.4]octane-6-carboxylate 1b
To a solution of iodolactone ( )-5 (0.51 g, 1.87 mmol) in DMF
(10 mL) was added with KOH (0.37 g, 6.60 mmol) and stirred for
24 h at room temperature. After removal of water in vacuo, the res-
idue was dissolved anhydrous DMF. The mixture was added EtI
(0.96 g, 6.56 mmol) at 40 °C, and stirred for 6 h. After removal of
volatile materials in vacuo, the reaction was quenched with NH4Cl
aq solution, and extracted with EtOAc (three times). The combined
organic phases were washed with brine and dried over Na2SO4, and

2110
M. Hamada et al. / Tetrahedron: Asymmetry 20 (2009) 2105–2111
NH2); 13C NMR (100 MHz, CDCl3): d = 29.6, 44.7, 62.8, 67.0, 67.7,
77.4, 78.1, 169.9, 170.4; IR mmax 3400, 3190, 2958, 1753, 1680,
1417, 1306, 1197 cmꢀ1. Anal. Calcd for C9H11NO5: C, 50.70; H,
5.20; N, 6.57. Found: C, 50.53; H, 5.15; N, 6.47.
Further recrystallization provided a sample of (+)-1a (49%
recovery, mp 112–113 °C), ½a D23
¼ þ15:8 (c 0.76, MeOH). This sam-
ꢁ
ple was revealed to be 96.4% ee by the HPLC analysis at the subse-
quent stage of 6a as below, and we concluded that the
enantiomeric excess of the acid 1a reaches constant value by rep-
etition of the recrystallization from EtOAc.
4.3.9. 2,2,2-Trifluoroethyl ( )-3,8-dioxatricyclo[3.2.1.02.4]-
octane-6-carboxylate 1g
The scaled-up experiment by applying ( )-1e (1.00 g, 4.59 mmol)
worked well in a reproducible manner to give (ꢀ)-1e: (230 mg,
A mixture of carboxylic acid ( )-1a (156 mg, 1.00 mmol), DMAP
(245 mg, 2.00 mmol), EDC-Cl (384 mg, 2.00 mmol), CF3CH2OH
(150 mg, 1.50 mmol), and triethylamine (202 mg, 2.00 mmol) in
DMF (1 mL) was stirred at room temperature under argon. The reac-
tion was monitored by silica gel TLC, developed with hexane–EtOAc
(1:4). After stirring for 10 h at room temperature, the mixture was
quenched by the addition of EtOAc–water. The organic materials
were extracted with EtOAc, and the combined organic phases were
washed with brine and dried over Na2SO4. The organic phase was
concentrated in vacuo. The residue was purified by silca gel column
chromatography with hexane–EtOAc (1:1) to afford trifluoroethyl
ester ( )-1g (192 mg, 80.6%) as a colorless oil; 1H NMR (270 MHz,
CDCl3): d = 1.93 (1H, ddd, J = 4.6, 11.3, 11.6 Hz, H-7exo), 2.04 (1H,
dd, J = 4.3, 11.6 Hz, H-7endo), 2.94 (1H, dt, J = 4.6, 11.3 Hz, H-6), 3.98
(1H, dd, J = 2.4, 4.6 Hz, H-4), 4.05 (1H, dd, J = 2.2, 4.6 Hz, H-2), 4.44
(1H, dddd, J = 8.4, 12.7 Hz, CH2CF3), 4.53 (1H, dt, J = 1.9, 4.9 Hz, H-
5), 4.58 (1H, dddd, J = 8.4, 12.7 Hz, CH2CF3), 4.69 (1H, dt, J = 1.9,
4.9 Hz, H-1); 13C NMR (100 MHz, CDCl3): d = 29.5, 44.4, 60.2, 60.6,
22.9%) ½a 2D4
¼ ꢀ5:3 (c 1.00, CHCl3); 99.7% ee after derivatization to
ꢁ
6b and its HPLC anlalysis. Acid (+)-1a: (428 mg, 59.8%)
½
a 2D4
sponding 6a.
ꢁ
¼ þ11:0 (c 1.00, MeOH); 77.3% ee by HPLC analysis of corre-
4.3.12. Methyl (1S,2R,4S,5R,6S)-(+)-3,8-dioxatricyclo[3.2.1.02.4]-
octane-6-carboxylate 1c
To a solution of the acid (+)-1a (30.6 mg, 0.20 mmol) as above in
anhydrous DMF was added Cs2CO3 (163 mg, 0.50 mmol) and CH3I
(85.1 mg, 0.60 mmol). The mixture was stirred at 50 °C for 24 h.
After concentration to dryness in vacuo, the residue was extracted
with EtOAc (three times), and the combined organic layer was
washed with brine, dried over Na2SO4, and concentrated in vacuo.
The residue was purified by preparative TLC with hexane–EtOAc
(1:1) to afford methyl ester (+)-1c (20.1 mg, 59%) as a colorless so-
lid. Mp 67–68 °C, ½a D23
¼ þ11:7 (c 0.57, MeOH). Its IR and NMR
ꢁ
spectra were identical with that of the authentic specimen.2
66.5, 66.6, 77.4, 77.5, 121.5, 124.3, 169.7; IR mmax 3010, 2969, 2368,
2337, 1747, 1411,1276, 1155, 879 cmꢀ1. HRMS (EI): calcd for
4.3.13. Methyl (1S,5S,6R)-(ꢀ)-5-hydroxy-7-oxabicyclo[4.1.0]-
C9H9F3O4: [M+]: 238.0453; found: m/z = 238.0453.
hept-2-en-3-carboxylate 6a
To
a solution of lithium hexamethyldisilazide [(TMS)2NLi,
4.3.10. PLE-catalyzed hydrolysis of esters 1b–1g
0.20 mL, 0.20 mmol] was added in THF (0.20 mL) at ꢀ78 °C. To a
solution of methyl ester (ꢀ)-1c (20.1 mg, 0.13 mmol) in THF
(0.20 mL) was added the LHMDS solution above dropwise at
ꢀ78 °C, and the mixture was stirred for 1 h at that temperature.
The reaction was quenched with saturated NH4Cl aq solution, and
extracted with EtOAc. The combined organic layer was washed with
brine, driedover Na2SO4, and concentratedin vacuo. Theresiduewas
purified by preparative TLC with hexane–EtOAc (1:1) to afford
methyl ester (ꢀ)-6a (14.0 mg, 70%, 90.6% ee) as a colorless solid.
The hydrolysis of each substrate was carried out under the same
conditions as described for the screening of enzymes with ethyl es-
ter 1b. The E-value of the each substrate was uniformly calculated
from the conversion and ee(P) as follows. The conversion was deter-
mined by 1H NMR analysis of crude reaction mixture. Ee(P) was
determined by the HPLC analysis at the stage of 6a, after methyla-
tion of hydrolyzate and following b-elimination as described later.
4.3.11. PLE-catalyzed hydrolysis of 2-chloroethyl ester 1e
To a stirred solution of 2-chloroethyl ester ( )-1e (373.3 mg,
1.71 mmol) in a phosphate buffer (0.2 M, pH 7.0; 8.5 mL), PLE (Sig-
½
a 2D3
ꢁ
¼ ꢀ200 (c 0.70, MeOH) [lit.:2
[a]D = +213 (c 0.56, MeOH), for
(1R,5R,6S)-6a]. The product (ꢀ)-6a was analyzed by HPLC [column,
Daicel Chiralcel OD-H, 0.46 cm ꢂ 25 cm; hexane–2-propanol (5:1);
flow rate 0.5 mL/min]: tR (min) = 15.1 (95.3%), 33.1 (4.7%).
ma, E2884, 850 lL) was added and the mixture was stirred for 24 h
at room temperature. The reaction was quenched with 1 M HCl to
pH 2, and extracted with EtOAc (10 times). The combined organic
layer was dried over Na2SO4 and concentrated in vacuo, and the ra-
tio between unreacted recovery 1e and hydrolyzate 1a was deter-
mined by 1H NMR measurement. The above mentioned crude
mixture was washed with saturated NaHCO3 aq solution. The or-
ganic layer was washed with brine and dried over Na2SO4, concen-
trated in vacuo to give (ꢀ)-1e (159.3 mg, 0.73 mmol) as the
unreacted recovery. The aqueous layer was acidified to pH 3 and
extracted with EtOAc (10 times). The extract was dried over
Na2SO4 and concentrated in vacuo to give (+)-1a (145.4 mg,
0.93 mmol, mp 107–108 °C). These samples were employed for
the next step without further purification.
Enantiomerically enriched acid (+)-1a (15.6 mg, 0.10 mmol) by
recrystallization in twice was treated with CH2N2 to give (+)-1c
(15.7 mg, 89%); mp 63–64 °C, ½a D23
¼ þ11:7 (c 0.75, MeOH). This
ꢁ
was converted to (ꢀ)-6a (11.1 mg, 74%, 96.4% ee); ½a D23
¼ ꢀ207 (c
ꢁ
0.55, MeOH). HPLC analysis was performed in the same manner:
tR (min) = 15.1 (98.2%), 33.1 (1.8%).
4.3.14. 2-Chloroethyl (1R,5R,6S)-(+)-5-hydroxy-7-oxabicyclo-
[4.1.0]hept-2-en-3-carboxylate 6b
In a similar manner as described for ( )-6b, 2-chloroethyl ester
(ꢀ)-1e (47.2 mg, 0.21 mmol) in THF (0.30 mL) was added with a
solution of lithium hexamethyldisilazide [(TMS)2NLi, 0.31 mL,
0.31 mmol] in THF (0.30 mL), gave (+)-6b (36.4 mg, 77%, 99.4%
Ester (ꢀ)-1e: ½a D23
¼ ꢀ5:3 (c 1.02, CHCl3), 99.4% ee as shown be-
ꢁ
ee: ½a 2D3
¼ þ233 (c 1.08, MeOH); The product (+)-6b was analyzed
ꢁ
low. Its IR and NMR spectra were in good accordance with those of
racemic sample. Acid (+)-1a: ½a D23
¼ þ11:7 (c 1.00, MeOH), 90.6% ee
ꢁ
by HPLC analysis [column, Daicel Chiralcel OD-H, 0.46 cm ꢂ 25 cm;
hexane–2-propanol (5:1); flow rate 0.5 mL/min]: tR (min) = 18.0
(0.3%), 38.0 (99.7%); 1H NMR (400 MHz, CDCl3): d = 2.32 (1H, ddd,
as shown below. This was further purified by recrystallization from
EtOAc to give (+)-1a (93.6 mg, 71%, mp 114–115 °C) ½a D23
¼ þ14:4
ꢁ
J = 3.3, 5.2, 17.6 Hz, H-6b), 2.80 (1H, dt, J = 2.1, 17.6 Hz, H-6a),
(c 0.75, MeOH). The sample obtained by recrystallization as above
(15.0 mg) was treated with CH2N2 to give (+)-1c (15.6 mg, 96%);
3.48 (1H, t, J = 3.9 Hz, H-3), 3.57 (1H, ddd, J = 2.1, 2.8, 3.9 Hz, H-
4), 3.69 (2H, t, J = 5.7 Hz, CH2Cl), 4.38 (2H, t, J = 5.7 Hz, CO2CH2),
4.57 (1H, br m, H-5), 7.19 (1H, dd, J = 3.3, 3.9 Hz, H-2); 13C NMR
(100 MHz, CDCl3): d = 29.3, 41.5, 46.2, 56.1, 63.5, 64.5, 130.3,
134.4, 165.5; IR mmax 3425, 2964, 1709, 1641, 1417, 1392, 1250,
mp 63–64 °C, ½a 2D3
¼ þ11:7 (c 0.75, MeOH). This was further con-
ꢁ
verted to (ꢀ)-6a (11.1 mg, 74%, 95.6% ee); ½a D23
¼ ꢀ207 (c 0.55,
ꢁ
MeOH). HPLC analysis was performed in the same manner: tR
(min) = 15.1 (97.8%), 33.1 (2.2%).

M. Hamada et al. / Tetrahedron: Asymmetry 20 (2009) 2105–2111
2111
1099, 918 cmꢀ1. Anal. Calcd for C9H11ClO4: C, 49.44; H, 5.07.
Found: C, 49.43; H, 5.36.
Scientific Research (No. 18580106) and ‘High-Tech Research Cen-
ter’ Project for Private Universities: matching fund subsidy 2006–
2011 from the Ministry of Education, Culture, Sports, Science and
Technology, Japan, and acknowledged with thanks.
4.3.15. (3S,4S,5R)-(ꢀ)-3,4,5-Trihydroxy-1-cyclohexene-1-
carboxylic acid 2
To a solution of (+)-6b (55.0 mg, 0.25 mmol) in THF and water
(1:1, 4 mL) was added KOH (21.2 mg, 0.38 mmol). After stirring
for 1 h at 50 °C, the mixture was neutralized with 1 M HCl to pH
3, and concentrated in vacuo. The solid was dissolved in water
References
1. Ogawa, S.; Yoshikawa, M.; Taki, T.; Yokoi, S.; Chida, N. Carbohydr. Res. 1995, 269,
53–78.
2. Shoji, M.; Imai, H.; Mukaida, M.; Sakai, K.; Kakeya, H.; Osada, H.; Hayashi, Y. J.
Org. Chem. 2005, 70, 79–91 and references cited therein.
3. Kakeya, H.; Miyake, Y.; Shoji, M.; Kishida, S.; Hayashi, Y.; Kataoka, T.; Osada, H.
Bioorg. Med. Chem. Lett. 2003, 13, 3743–3746.
4. Evans, D. A.; Barnes, D. M. Tetrahedron Lett. 1997, 38, 57–58.
5. Sarakinos, G.; Corey, E. J. Org. Lett. 1999, 1, 1741–1744.
6. Ogawa, S.; Iwasawa, Y.; Nose, T.; Suami, T.; Ohba, S.; Ito, M.; Saito, Y. J. Chem.
Soc., Perkin Trans. 1 1985, 903–906 and references cited therein.
7. Kurokawa, M.; Sugai, T. Bull. Chem. Soc. Jpn. 2004, 77, 1021–1025.
8. Tran, C. H.; Crout, D. H. G. J. Chem. Soc., Perkin Trans. 1 1998, 1065–1068.
9. Fotakopoulou, I.; Barbayianni, E.; Constantinou-Kokotou, V.; Bornscheuer, U. T.;
Kokotos, G. J. Org. Chem. 2007, 72, 782–786 and references cited therein.
10. Hirose, Y.; Kariya, K.; Sasaki, I.; Kurono, Y.; Achiwa, K. Tetrahedron Lett. 1993,
34, 5915–5918.
(1 mL) and trifluoroacetic acid (400 lL, 5.39 mmol) was added to
the solution with stirring. The mixture was stirred for 3 h at
50 °C. The reaction mixture was concentrated in vacuo to remove
volatile materials. The residue was purified by silica gel column
chromatography with CHCl3–MeOH (10:1) to afford carboxylic acid
(ꢀ)-2: ½a 2D5
ꢁ
¼ ꢀ33:1 (c 0.34, H2O) [lit.:20
[a
]
D = ꢀ31.0 (c 0.1, H2O)];
1H NMR (400 MHz, D2O): d = 2.06 (1H, dddd, J = 2.8, 4.0, 10.0,
16.8 Hz, H-6 b), 2.61 (1H, ddd, J = 1.6, 6.0, 16.8 Hz, H-6a), 3.33
(1H, dd, J = 8.4, 10.0 Hz, H-4), 3.62 (1H, dt, J = 6.0, 10.0, 10.0 Hz,
H-5), 4.11 (1H, dddd, J = 1.6, 2.4, 4.0, 8.4 Hz, H-3), 6.51 (1H, dd,
J = 2.4, 2.8 Hz, H-2); 13C NMR (100 MHz, D2O): d = 31.7, 68.6,
71.5, 76.2, 128.2, 139.2, 169.7; IR mmax 3261, 1556, 1409, 1072
cmꢀ1. Its 1H NMR spectrum was identical with that reported previ-
ously.21 As this product 2 is a trihydroxy acid and shows highly
hydrophilic property and is susceptible to an irreversible adsorp-
tion on silica gel, the yield was estimated to be 79.7%, at the stage
just before the final purification, based on 1H NMR with an internal
11. Provencher, L.; Jones, J. B. J. Org. Chem. 1994, 59, 2729–2732.
12. Terashima, S.; Ujita, K.; Kuwayama, T.; Sugioka, T.; Shimizu, K.; Kanehira, K.
WO 2000-JP8348 20001127 (CA 135:76775), 2001. They prepared racemic
oseltamivir phosphate from ( )-1a.
13. Abrecht, S.; Harrington, P.; Iding, H.; Karpf, M.; Trussardi, R.; Wirz, B.; Zutter, U.
Chimia 2004, 58, 621–629.
14. Shibasaki, M.; Kanai, M. Eur. J. Org. Chem. 2008, 1839–1850 and references cited
therein.
15. Darths, K. M.; Knop, D. R.; Frost, J. W. J. Am. Chem. Soc. 1999, 121, 1603–1604.
16. Adachi, O.; Ano, Y.; Toyama, T.; Matsushita, K. Biosci., Biotechnol., Biochem.
2006, 70, 2579–2582.
standard [methyl b-D-glucoside, Tokyo Kasei Co., M709, analyti-
cally pure grade, standard signal at d = 4.23 (1H, d, J = 7.6 Hz)].
17. Sánchez-Abella, L.; Fernández, S.; Verstuyf, A.; Verlinden, L.; Ferrero, M.; Gotor,
V. Bioorg. Med. Chem. 2007, 15, 4193–4202.
18. Zhang, Y.; Liu, A.; Ye, Z. G.; Lin, J.; Xu, L. Z.; Yang, S. L. Chem. Pharm. Bull. 2006,
54, 1459–1461.
Acknowledgements
The authors thank Professor Kaoru Nakamura of Institute for
Chemical Research, Kyoto University, for his valuable discussion
and encouragement on this study. Meito Sangyo Co. Ltd and Nag-
ase ChemteX Co. were acknowledged with thanks, for generous gift
of enzymes. This work was supported both by a Grant-in-Aid for
19. Huang, J.; Chen, F.-E. Helv. Chim. Acta 2007, 90, 1366–1372.
20. Tan, D. S.; Foley, M. A.; Shair, M. D.; Schreiber, S. L. J. Am. Chem. Soc. 1998, 120,
8565–8566.
21. Brettle, R.; Cross, R.; Frederickson, M.; Haslam, E.; MacBeath, F. S.; Davies, G. M.
Tetrahedron 1996, 52, 10547–10556.